AI Glossary
The complete dictionary of Artificial Intelligence
Volumetric Segmentation
Process of partitioning a three-dimensional space into distinct regions, aiming to assign a semantic label to each voxel of a data volume.
Polygon Mesh
Representation of a 3D object by a collection of vertices, edges, and faces (polygons), forming a continuous surface.
Signed Distance Field (SDF)
Function that, for each point in space, returns the distance to the nearest surface, with the sign indicating whether the point is inside or outside the object.
Point Cloud Fusion
Technique involving combining multiple point clouds, often from different perspectives or capture sessions, to create a single, complete 3D model.
Octree
Tree data structure used to partition a 3D space by recursively subdividing a cube into eight sub-cubes, optimizing the processing and rendering of volumetric data.
3D Graph Neural Network (GNN)
Neural network applied to graph-structured data, where nodes represent points in a point cloud and edges encode spatial relationships.
Computed Tomography (CT Scan)
Medical imaging technique that reconstructs a 3D volume from a series of X-rays taken at different angles, producing a cross-sectional view of the object.
Magnetic Resonance Imaging (MRI)
Non-invasive medical imaging procedure using a magnetic field and radio waves to generate detailed 3D images of the inside of the body.
3D Instance Segmentation
Computer vision task that not only classifies each voxel but also distinguishes each object occurrence within the same semantic class in a volume.
3D U-Net
Convolutional neural network architecture adapted for volumetric segmentation, characterized by an encoder-decoder structure with skip connections to preserve spatial details.
V-Net
Variant of 3D U-Net that replaces standard convolutions with residual convolutions, improving gradient propagation and performance on medical volumes.
Query Sphere
In the context of point clouds, spherical volume defined around a point to collect its local neighbors and compute geometric features.
Level Set
Numerical method for tracking interfaces and shapes, representing a surface as the zero level of a higher-dimensional function.
Occupancy Function
Function that associates each point in 3D space with a probability or binary value indicating whether the space is occupied by an object or empty.